Llama 3.1 70B Instruct [4bit]: benchmark results
Provider: Meta.
Aggregate Rating 1553 ± 50, rank #1666 of 10916 rated models, from 17 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open Portuguese LLM - FaQuAD NLI | 81.69 | Macro F1 (%) | 92.9 |
| Open Portuguese LLM - BLUEX | 70.65 | Accuracy (%) | 92.2 |
| Open Portuguese LLM - OAB Exams | 58.91 | Accuracy (%) | 91.6 |
| Open Portuguese LLM - ENEM | 74.74 | Accuracy (%) | 87.3 |
| Open Portuguese LLM - ASSIN2 RTE | 92.64 | Macro F1 (%) | 81.4 |
| Czech Opinion ABSA (10-Shot) - ASQP | 40.99 | Micro F1 (%) of exact-match aspect sentiment quad (aspect te | 77.8 |
| Czech Opinion ABSA (10-Shot) - ACOS | 33.37 | Micro F1 (%) of exact-match aspect-category-opinion-sentimen | 66.7 |
| Czech Opinion ABSA (Zero-Shot) - ASQP | 31.86 | Micro F1 (%) of exact-match aspect sentiment quad (aspect te | 66.7 |
| Czech Opinion ABSA (10-Shot) | 40.86 | Mean of the ASTE, ASQP and ACOS micro F1 scores (%) on the C | 55.6 |
| Czech Opinion ABSA (10-Shot) - ASTE | 48.23 | Micro F1 (%) of exact-match aspect sentiment triplet (aspect | 55.6 |
| Czech Opinion ABSA (Zero-Shot) | 34.76 | Mean of the ASTE, ASQP and ACOS micro F1 scores (%) on the C | 55.6 |
| Czech Opinion ABSA (Zero-Shot) - ACOS | 26.49 | Micro F1 (%) of exact-match aspect-category-opinion-sentimen | 55.6 |
Interactive version: theaggregate.ai/model?slug=llama-3-1-70b-instruct-4bit · How It Works · Data refreshed daily, snapshot 2026-11-08.